Post by thread:[EE] AN521, A Parting Shot

I can't drop this, it is too important. I have never seen an app note like this- sans safety warnings- anywhere else in the industry. What we experienced designers are concerned about is that AN51- if followed exactly- has the potential to cause a product to be created that might hurt someone. People look to MicroChip, a world leader, to provide guidance. Since the word "APPLICATION" is used in the title, people without much engineering knowledge will assume that what is said here is safe, 'cause a world leader said so. I would not be very concerned if this note come from a user, but coming from an actual _MicroChip_ _employee_, it opens MicroChip up to a needless lawsuit if the note is followed without serious revision. Microchip should: 1. State a clear warning to the reader that the circuit should be isolated with an isolation transformer before working with any design powered straight from the power line. 2. Indicate that the limiting resistor should be constructed so that it will handle AC line transients. What is most damning is that the contrary is indicated by the app note...!! 3. Indicate that the PCB layouts must be carefully done to prevent a creepage path that might harm a user. OR Remove AN251 and all references to it from its catalog. --Bob -- Note: To protect our network, attachments must be sent to spam_OUTattachTakeThisOuTengineer.cotse.net . 1-520-777-7606 USA/Canada http://beam.to/azengieer

are you talking about the 110vac to 5vac transformer eliminator? or the wheatstone bridge adc? you gave two different numbers, 51 and 251, I think the the transformer eliminator is TB008 or something. On 10/28/05, Bob Axtell <.....engineerKILLspam@spam@cotse.net> wrote: {Quote hidden}> > I can't drop this, it is too important. I have never seen an > app note like this- sans safety warnings- anywhere else > in the industry. > > What we experienced designers are concerned about is > that AN51- if followed exactly- has the potential to cause > a product to be created that might hurt someone. > > People look to MicroChip, a world leader, to provide guidance. > Since the word "APPLICATION" is used in the title, people > without much engineering knowledge will assume that what > is said here is safe, 'cause a world leader said so. I would > not be very concerned if this note come from a user, but > coming from an actual _MicroChip_ _employee_, > it opens MicroChip up to a needless lawsuit if the note > is followed without serious revision. > > Microchip should: > > 1. State a clear warning to the reader that the circuit should be > isolated with an isolation transformer before working with any design > powered straight from the power line. > > 2. Indicate that the limiting resistor should be constructed so that > it will handle AC line transients. What is most damning is that > the contrary is indicated by the app note...!! > > 3. Indicate that the PCB layouts must be carefully done to prevent > a creepage path that might harm a user. > > OR > > Remove AN251 and all references to it from its catalog. > > --Bob > > -- > Note: To protect our network, > attachments must be sent to > attachKILLspamengineer.cotse.net . > 1-520-777-7606 USA/Canada > http://beam.to/azengieer > >

Sorry, AN521. John Remington wrote: {Quote hidden}>are you talking about the 110vac to 5vac transformer eliminator? or the >wheatstone bridge adc? > >you gave two different numbers, 51 and 251, I think the the transformer >eliminator is TB008 or something. > > > > > >On 10/28/05, Bob Axtell <.....engineerKILLspam.....cotse.net> wrote: > > >>I can't drop this, it is too important. I have never seen an >>app note like this- sans safety warnings- anywhere else >>in the industry. >> >>What we experienced designers are concerned about is >>that AN51- if followed exactly- has the potential to cause >>a product to be created that might hurt someone. >> >>People look to MicroChip, a world leader, to provide guidance. >>Since the word "APPLICATION" is used in the title, people >>without much engineering knowledge will assume that what >>is said here is safe, 'cause a world leader said so. I would >>not be very concerned if this note come from a user, but >>coming from an actual _MicroChip_ _employee_, >>it opens MicroChip up to a needless lawsuit if the note >>is followed without serious revision. >> >>Microchip should: >> >>1. State a clear warning to the reader that the circuit should be >>isolated with an isolation transformer before working with any design >>powered straight from the power line. >> >>2. Indicate that the limiting resistor should be constructed so that >>it will handle AC line transients. What is most damning is that >>the contrary is indicated by the app note...!! >> >>3. Indicate that the PCB layouts must be carefully done to prevent >>a creepage path that might harm a user. >> >>OR >> >>Remove AN251 and all references to it from its catalog. >> >>--Bob >> >>-- >>Note: To protect our network, >>attachments must be sent to >>EraseMEattachspam_OUTTakeThisOuTengineer.cotse.net . >>1-520-777-7606 USA/Canada >>http://beam.to/azengieer >> >>-

> Microchip should: > > 1. State a clear warning to the reader that the circuit should be > isolated with an isolation transformer before working with any design > powered straight from the power line. > > 2. Indicate that the limiting resistor should be constructed so that > it will handle AC line transients. What is most damning is that > the contrary is indicated by the app note...!! > > 3. Indicate that the PCB layouts must be carefully done to prevent > a creepage path that might harm a user. An article in the latest Silicon Chip, who are generally very sensible, has an electric fan controller based on the F88. Mains cycles for timing are injected into RA3 via a series 1.5M VR25 resistor (2 or 3 of these in series would be better as the mains around here is 240V. Current into RA3 will be around 235uA). RA3 has a 4n7 polyester cap to Vss Although the article has several warnings about lethality in bold print, they also neglect to mention an isolating transformer. I can see someone losing a scope. The plastic timer trimpot they specify could have been better. A larger one with a plastic knob for example. And it's left up to the constructor to label the unit as dangerous.............

Jinx wrote: {Quote hidden}>>Microchip should: >> >>1. State a clear warning to the reader that the circuit should be >>isolated with an isolation transformer before working with any design >>powered straight from the power line. >> >>2. Indicate that the limiting resistor should be constructed so that >>it will handle AC line transients. What is most damning is that >>the contrary is indicated by the app note...!! >> >>3. Indicate that the PCB layouts must be carefully done to prevent >>a creepage path that might harm a user. >> >> > >An article in the latest Silicon Chip, who are generally very sensible, >has an electric fan controller based on the F88. Mains cycles for >timing are injected into RA3 via a series 1.5M VR25 resistor (2 >or 3 of these in series would be better as the mains around here is >240V. Current into RA3 will be around 235uA). RA3 has a 4n7 >polyester cap to Vss > >Although the article has several warnings about lethality in bold >print, they also neglect to mention an isolating transformer. I can >see someone losing a scope. The plastic timer trimpot they >specify could have been better. A larger one with a plastic knob >for example. And it's left up to the constructor to label the unit >as dangerous............. > > > I myself have designed a commercial product with mains to 5VDC supply, and an RC scheme to provide mains 60hz timing. But when it was being developed, I used an isolation transformer. The design contained a low-value line fuse, an AC spike suppressor (UL rated) and when it was tested in production, was potted in a black potting box. The timing resistor was TWO 1/2w resistors in series, and NO signals ran under these resistors. Our mains is 115VAC 60Hz, so the creepage path needs are lower than 240VAC. None have ever failed, they are still working 3 years later. --Bob -- Note: To protect our network, attachments must be sent to attachspam_OUTengineer.cotse.net . 1-520-777-7606 USA/Canada http://beam.to/azengineer

> >Although the article has several warnings about lethality in bold > >print, they also neglect to mention an isolating transformer. I can > >see someone losing a scope. The plastic timer trimpot they > >specify could have been better. A larger one with a plastic knob > >for example. And it's left up to the constructor to label the unit > >as dangerous............. > I myself have designed a commercial product with mains to 5VDC supply, and > an RC scheme to provide mains 60hz timing. But when it was being > developed, > I used an isolation transformer. As someone who has always lived almost entirely in tiny voltage DC land, I feel I need to ask: What do you mean by "isolation transformer"? Are you suggesting that ALL measurements on a system which is not galvanically isolated from the mains should be done through a 1:1 transformer? Mike H.

> Are you suggesting that ALL measurements on a system > which is not galvanically isolated from the mains should be > done through a 1:1 transformer ? No problem using a multimeter for most signals, and you don't *need* to isolate a circuit for that. But often you'd want to know the quality of those signals, and a scope is going to be the common tool for that. You can't just hook up scope ground to the Vss of such a circuit, so either the circuit (or the scope, not so likely, depends on the scale of what needs isolating) has to be powered through an isolating transformer. Vss becomes E when you earth it with the scope. You still don't want to be poking your fingers where they aren't welcome though, plenty of smoke and screams itching to be let loose

At 11:39 AM 11/1/2005 +1300, you wrote: > > Are you suggesting that ALL measurements on a system > > which is not galvanically isolated from the mains should be > > done through a 1:1 transformer ? > >No problem using a multimeter for most signals, and you >don't *need* to isolate a circuit for that. But often you'd want >to know the quality of those signals, and a scope is going to >be the common tool for that. You can't just hook up scope >ground to the Vss of such a circuit, so either the circuit (or the >scope, not so likely, depends on the scale of what needs >isolating) has to be powered through an isolating transformer. Some folks have been known to perform a pinectomy on the power supply cord of the scope. I strongly advise against this course of action, as it can leave the scope and associated controls energized at line voltage. And thus result in shocking, sizzling, funerals, good stuff being sold for a song, friends comforting your bereaved spouse in generally unacceptable ways, etc. etc. >Vss becomes E when you earth it with the scope. You still >don't want to be poking your fingers where they aren't welcome >though, plenty of smoke and screams itching to be let loose Using an emulator or ISP on the system when one side is tied to the line and one side of your computer is earthed can lead to a lot of expensive smoke. Also, with one side tied to the line you're *one* slip rather than *two* away from serious damage or injury. Normally the isolation transformer would go between the power supply and the system under test. You still have the high voltage issues, but you are safer. Personally, I like to build safe isolation into the power supply and the output stage etc. (eg. I/O optoisolators mounted securely in a separate box, transformer isolated power supply ) for testing the firmware in such systems. It's way too easy to have a minor problem cause a lot of damage-- it only takes a wee little mistake. Best regards, Spehro Pefhany --"it's the network..." "The Journey is the reward" @spam@speffKILLspaminterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com ->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff

Mike Hord wrote: {Quote hidden}>>>Although the article has several warnings about lethality in bold >>>print, they also neglect to mention an isolating transformer. I can >>>see someone losing a scope. The plastic timer trimpot they >>>specify could have been better. A larger one with a plastic knob >>>for example. And it's left up to the constructor to label the unit >>>as dangerous............. >>> >>> > > > >>I myself have designed a commercial product with mains to 5VDC supply, and >>an RC scheme to provide mains 60hz timing. But when it was being >>developed, >>I used an isolation transformer. >> >> > >As someone who has always lived almost entirely in tiny voltage >DC land, I feel I need to ask: > >What do you mean by "isolation transformer"? Are you suggesting >that ALL measurements on a system which is not galvanically >isolated from the mains should be done through a 1:1 transformer? > >Mike H. > > > Yes, absolutely, because you never know when you might have a break in the return path, and make everything hot at 115VAC. --Bob -- Note: To protect our network, attachments must be sent to KILLspamattachKILLspamengineer.cotse.net . 1-520-777-7606 USA/Canada http://beam.to/azengineer

>> I used an isolation transformer. > > As someone who has always lived almost entirely in tiny voltage > DC land, I feel I need to ask: > > What do you mean by "isolation transformer"? Are you suggesting > that ALL measurements on a system which is not galvanically > isolated from the mains should be done through a 1:1 transformer? Hi Mike, Actually the isolation transformer is a 1:1 power transformer, so the line power to your circuit is isolated and no part of the circuit is "live" with reference to ground. That way you don't have to "float your scope" with the potential of making the scope's case live. etc. (Although you also can use a fully isolated scope like my THS720 but to be fully safe you'd want to use a safety probe which has better insulation on the ground clip. But that doesn't protect you from shock hazard if YOU touch part of the circuit.) Cheers, Bob

>> Are you suggesting that ALL measurements on a system >> which is not galvanically isolated from the mains should be >> done through a 1:1 transformer ? > No problem using a multimeter for most signals, and you > don't *need* to isolate a circuit for that. ***** WARNING **** ***** WARNING **** ***** WARNING **** Multimeters do have voltage ratings, both for isolation and for maximum applied voltages. While this is usually not an issue, it can be. El cheapo meter leads may have an exposed metallic path to the lead - very naughty and potentially fatal. Many bottom end meters are not rated for 240 VAC mains use even though they have a range that displays full mains voltage. Battery door needs to be secured in place to avoid possible unexpected results when HV is applied. I have had a meter set to its 1000 VDC range erupt in my hands when measuring what turned out to be slightly over 1000 VDC (probably about 1200 V). This was an analog meter with a PCB switch. Breakdown occurred across the switch in some manner. Most exciting. RM